The goals of this project are to model the binding of a selection of antitumor drugs to double helical DNA using molecular mechanics and computer graphics. Where it is appropriate, attempts to verify these models by physical measurements including 2D- NOESY experiments will be undertaken. The resulting models should be valuable in understanding the mode of action of these drugs and in the subsequent rational design of their analogs. Specific aims are as follows. 1) Derive a binding model for sibiromycin, a pyrrolo(1,4)benzodiazepine bearing an aminosugar. Explore the use of molecular mechanics in explaining the preference of neothramycin A for poly(dG) . poly(dC) over poly (dGdC) . poly(dGdC). 2) Model the binding of saframycin A, naphthyridinomycin, and quinocarcin, examples of three types of drugs that appear to act by a similar process, groove binding followed by alkylation. Verify this binding process by 13C-NMR and 2D-NMR experiments. 3) Evaluate the possibility that streptonigrin is a groove binder by appropriate physical measurements. 4) Develop a model for the intercalative binding of ellipticene and important analogs. Also model the binding of the o-quinone derived from 9-hydroxyellipticene and see if it can react with a nucleophile on the DNA while bound. 5) Investigate the intercalation of leucanthone and analogs by deriving a computer model and comparing it with the 2D-NMR spectrum of a DNA complex. Also, compare Tm measurements on DNA, free and bound to these compounds, with calculated relative binding strengths. The 2D-NMR experiments will include NOESY and COSY techniques to locate exact binding interactions between atoms in the drugs and in the oligonucleotides d(ATGCAT)2 and d(GCATATGC)2. Other 1H-NMR experiments will include T2 measurements to probe structural and dynamic aspects of the adducts and 1D-NOE measurements to assign adenine H(2) protons in the oligonucleotides and their adducts.